4 upvotes
OpenFlyV2 lets drones follow language commands in flight
What is OpenFlyV2: An onboard vision-language-action model for aerial perception and navigation?
OpenFlyV2 is a vision-language-action model that runs on a drone’s onboard computer. It translates spoken or written instructions and live camera footage into navigation actions without needing a ground-station.
Who it is for
Robotics developers can now prototype language-guided aerial tasks without building separate perception and control stacks.
How to try it
Visit the model card on Hugging Face and run the provided demo notebook